Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Particle filtering in Computer Vision (2003)

1,597 views

Published on

Published in: Education, Technology, Business
  • Be the first to comment

Particle filtering in Computer Vision (2003)

  1. 1. Particle Filtering Ph.D. Coursework: Computer Vision Eric Lehmann Department of Telecommunications EngineeringResearch School of Information Sciences and Engineering The Australian National University, Canberra Eric.Lehmann@anu.edu.au June 06, 2003
  2. 2. ContextTracking:• Probabilistic inference about the motion of an object given a sequence of measurements• Applications: robotics, multimedia, military, videoconferencing, surveillance, etc.• Computer vision: vehicle tracking, human-computer interaction, robot localisation, etc.In practice:• Noise in measurements (images)• Background might be heavily clutteredÉ Robust tracking method: state-space approachPage 1 — Computer Vision, Ph.D. coursework Eric Lehmann, Tel. Eng.
  3. 3. State-Space ApproachProblem definitions:• State variable X k : e.g. target position and velocity in state-space at time k X k = [x, y, z, x, y, z]T ˙ ˙ ˙• Observation Y k : measurements obtained from processing camera image data• Set of all observations: Y 1:k = [Y 1, . . . , Y k ]• System dynamics (transition) equation: X k = g(X k−1, v k−1)Aim: given all data Y 1:k , compute posterior PDF p(X k |Y 1:k ) É Bayesian filtering problemPage 2 — Computer Vision, Ph.D. coursework Eric Lehmann, Tel. Eng.
  4. 4. State-Space Approach• Bayesian filtering solution: if posterior PDF p(X k−1|Y 1:k−1) known at time k − 1, compute current posterior PDF as follows: Predict: p(X k |Y 1:k−1) = p(X k |X k−1) p(X k−1|Y 1:k−1) dX k−1 Update: p(X k |Y 1:k ) ∝ p(Y k |X k ) p(X k |Y 1:k−1) where p(Y k |X k ) is the likelihood function (measurement PDF)• Problem: usually no closed-form solutions available for many natural dynamic models• Current approximations: Kalman filter, extended Kalman filter, Gaussian sum methods, grid-based methods, etc. É Sequential Monte Carlo methods, i.e. Particle Filters (PF)Page 3 — Computer Vision, Ph.D. coursework Eric Lehmann, Tel. Eng.
  5. 5. State-Space Approach: Symbolic RepresentationCase: Gaussian noise and linear equationsFrom [Condensation – conditional density propagation for visual tracking, Isard and Blake, Int. J.Computer Vision, 1998]Page 4 — Computer Vision, Ph.D. coursework Eric Lehmann, Tel. Eng.
  6. 6. State-Space Approach: Symbolic RepresentationCase: non-Gaussian noise and/or nonlinear equationsFrom [Condensation – conditional density propagation for visual tracking, Isard and Blake, Int. J.Computer Vision, 1998]Page 5 — Computer Vision, Ph.D. coursework Eric Lehmann, Tel. Eng.
  7. 7. Particle Filtering• Numerical method to solve nonlinear and/or non-Gaussian Bayesian filtering problems• Known variously as: bootstrap filtering, condensation algorithm, interacting particle approximations, survival of the fittest, JetStream, etc.• Particle and weight representation of posterior density:From [Condensation – conditional density propagation for visual tracking, Isard and Blake, Int. J.Computer Vision, 1998]Page 6 — Computer Vision, Ph.D. coursework Eric Lehmann, Tel. Eng.
  8. 8. Basic PF AlgorithmFrom [Novel approach to nonlinear/non-Gaussian Bayesian state estimation, Gordon et al., IEEProc. F., 1993]Assumption: a set of N state samples and corresponding weights (i) (i){X k−1, wk−1, i = 1, . . . , N } represents the posterior densityp(X k−1|Y 1:k−1) at time k − 1Procedure: update the particle set to represent the posterior densityp(X k |Y 1:k ) for current time k according to following iterationsPage 7 — Computer Vision, Ph.D. coursework Eric Lehmann, Tel. Eng.
  9. 9. Basic PF: Symbolic Representation (i) (i) {X k−1, wk−1} ∼ p(X k−1|Y 1:k−1) ⇐ resampling (i) {X k−1, 1/N } ∼ p(X k−1|Y 1:k−1) ⇐ prediction (i) {X k , 1/N } ∼ p(X k |Y 1:k−1)p(Y k |X k ) ⇐ measurement & update Xk (i) (i) {X k , wk } ∼ p(X k |Y 1:k )Page 8 — Computer Vision, Ph.D. coursework Eric Lehmann, Tel. Eng.
  10. 10. PF Methods Overview• Algorithm design choices: Source dynamics model: various models available Observations: camera image data Likelihood function: derived from observations• Large number of enhanced PF versions to be found in literature: auxiliary PF, unscented PF, ICondensation, hybrid bootstrap, fast weighted bootstrap, annealed PF, etc.• PF methods: special case of Sequential Importance Sampling, see [On sequential Monte Carlo sampling methods for Bayesian filtering, Doucet et al., Statist. Comput., 2000]• Excellent review of current PF methods in [A tutorial on particle filters for online nonlinear/non-Gaussian Bayesian Tracking, Arulampalam et al., IEEE Trans. Sig. Proc., 2002]Page 9 — Computer Vision, Ph.D. coursework Eric Lehmann, Tel. Eng.
  11. 11. PF Tracking of a Head OutlineStandard head outline template (parametric spline curve) used fortracking. Measurements are obtained by detecting maxima ofintensity gradient along lines normal to the head contour.From [Condensation – conditional density propagation for visual tracking, Isard and Blake, Int. J.Computer Vision, 1998] and [Sequential Monte Carlo fusion of sound and vision for speakertracking, Vermaak et al., Proc. Int. Conf. on Computer Vision, 2001]Page 10 — Computer Vision, Ph.D. coursework Eric Lehmann, Tel. Eng.
  12. 12. PF Tracking of a Head OutlineParticle representation of shape distributionFrom [Condensation – conditional density propagation for visual tracking, Isard and Blake, Int. J.Computer Vision, 1998]Page 11 — Computer Vision, Ph.D. coursework Eric Lehmann, Tel. Eng.
  13. 13. Application ExampleTracking objects in heavy clutter hand.mpg dancemv.mpg leafmv.mpgFrom [Condensation – conditional density propagation for visual tracking, Isard and Blake, Int. J.Computer Vision, 1998]Page 12 — Computer Vision, Ph.D. coursework Eric Lehmann, Tel. Eng.
  14. 14. Application ExampleCombining sound and vision in PF algorithm pat jacoC out.aviFrom [Sequential Monte Carlo fusion of sound and vision for speaker tracking, Vermaak et al.,Proc. Int. Conf. on Computer Vision, 2001]Page 13 — Computer Vision, Ph.D. coursework Eric Lehmann, Tel. Eng.
  15. 15. Application ExampleTracking of more complex models walker.mpgFrom [Articulated Body Motion Capture by Annealed Particle Filtering, Deutscher et al., IEEEConf. Computer Vision and Pattern Recognition, 2000]Page 14 — Computer Vision, Ph.D. coursework Eric Lehmann, Tel. Eng.

×